Still (as of May, 2009) one of my few favorite psychology treatises of all time, A Theory of Cognitive Dissonance best exemplified the pivotal concept of, and was a key bibliographic source for, my 1972 Georgetown Univ. doctoral dissertation topic (the title itself was longer, applied to specific subjects), "Cognitive Consistency and Interpersonal Conflict."
I'm revisiting this subject today because just two days ago a local professional psychologist I met at the gym described Festinger's book as "maybe the highest overall explanatory power regarding human behavior of anything I've ever read." The gist of Festinger's theory, distilled from his ample research and experience, is that there is systemic pressure toward organic consistency (the opposite of dissonance) among each person's percepts, concepts, overall sociological interactions and worldview tendencies, and that new information felt to be inconsistent with those is strongly resisted, often not understood or even clearly perceived because this new info is dissonant, i.e. cacophonous, with the person's already incorporated data/feelings/beliefs.
Since 1972 I've found nothing to disconfirm Festinger's theory, and dozens of elements to confirm it. One recent excellent lay presentation, the cover article of New Scientist magazine of about a year ago, "Two Tribes," summarizes two genetically distinct brain types that greatly differ, for example, in a standard cognitive measurement known as "tolerance of ambiguity." The article, directed mainly toward Americans, presented tested data demonstrating that, in general, individuals of higher tolerance of ambiguity were more frequently found among Democrats, and those of lower tolerance more often among Republicans. Individuals of lower ambiguity tolerance find it harder to incorporate new info and reassess/rearrange prior info, because they experience more physical/mental stress during periods of shades-of-gray ambiguity; their brain more quickly pulls together data perceived as consistent with prior beliefs and more quickly excludes/rejects anything dissonant with it. (As a prime example of low ambiguity tolerance, think of GW Bush, who often called himself "the decider.") States of indecision are much more painful for the low-tolerance type, who then hold tighter to whatever opinions they form, even in the face of clearly disconfirming/falsifying information.
As illustrative of this powerfully confirmed theory, Festinger also wrote When Prophecy Fails, the true description of how a group of religious believers in a contemporary end-of-world prophet reacted when the prophet's specifically promised world-end prediction glaringly failed on the guaranteed date. Rather than sensibly seeing the failure as dissonant proof of a false prophet, his believers rationalized the non-end in a way consistent with their preferred belief structure. His book Theory of Cognitive Dissonance explains how this psychology operates as the result of organic brain function that can only put together "pictures" for us that we can make sense of, and "making sense" can mainly be done by linking new input to what we already believe. This is also the principle underlying all optical, aural and tactile illusions, even phantom-limb syndrome (which Festinger doesn't discuss) which results from the brain's "sense-making" function being unable to incorporate the perceived dissonance of an absent limb.
I regard it as somewhat tragic that A Theory of Cognitive Dissonance is out of print, since I think it's abundantly clear that it provides an enlightening explanatory model akin to what Charles Darwin's book, On the Origin of Species, accomplished for biology.